DE60208224T2 - MAGNETORESISTIVE HIGH-PERFORMANCE SPIN VALVE ASSEMBLY - Google Patents

Abstract

Magnetoresistive device comprising a spin valve formed from a stack of layers including at least two magnetic layers for which a relative orientation of their magnetisation directions are capable varying under influence of a magnetic field; at least one discontinuous dielectric or semiconducting layer with electrically conducting bridges at least partially passing through a thickness of the dielectric or semiconducting layer, the bridges configured to locally concentrate current that passes transversely through the stack; and means for circulating a current in the spin valve transverse to the plane of the layers, characterised in that the dielectric or semiconducting layer with electrically conducting bridges is arranged inside one of the magnetic layers.

Description

TECHNICAL TERRITORY

The The present invention relates to a magnetoresistive spin valve device or arrangement. This arrangement may be in the field of ultra-dense magnetic Recording can be used as a sensitive element in the magnetoresistive read heads of magnetic tapes or from computer hard drives. This arrangement can also be realized memory points and memories of type MRAM (for "Magnetic Random Access Memory or Magnetic Main memory).

STATE OF THE ART

In the field of magnetic read heads as well as the magnetic random access memory, magnetoresistive spin valve arrangements have recently been used whose spin valve is formed by a stack of layers having a sandwich layer sandwiched between two magnetic layers. The relative orientation of the magnetization directions of these layers may vary under the influence of a magnetic field. In certain arrangements, the two magnetic layers have a free direction of magnetization. One then uses the two extreme positions or constellations, where the two magnetization directions are either parallel or antiparallel, but also all intermediate positions. In other arrangements, the direction of magnetization is free in one of the magnetic layers and not free in the other. The direction of their magnetization is determined by an antiferromagnetic exchange layer. The American patent US 5,898,548 shows such a configuration wherein the separation layer is a thin nonconductive barrier forming a tunnel junction.

the Spin valves are added to facilities that allow a current to flow vertically flow to the level of the layers allow. They have the form of electrodes on the top and bottom of the spin valve.

There are also arrangements in which the current flows perpendicular to the plane of the layers and whose spin valve is completely metallic, with a continuous metal layer 10 sandwiched between two magnetic layers 20 . 30 arranged as in the American patent US 5,688,688 , Such a device is in the 1 shown. The magnetic free layer 30 and the magnetically fixed layer 20 be through a continuous separation layer 10 separated, which is electrically conductive and non-magnetic to the two magnetic layers 20 . 30 to decouple. On both sides of the magnetic layers 20 . 30 are power supply electrodes 60a . 60b shown.

In this example, the magnetically fixed layer is 20 with an antiferromagnetic layer 21 connected, which helps to maintain their magnetization. It is located between the magnetically fixed layer 20 and the associated electrode 60a , The magnetically free layer 30 is a buffer layer 31 assigned (known under the name of "buffer layer"), which helps to promote the growth of all the spin valve forming layers.

These The latter arrangements are often GMR arrangements (for Giant MagntoResistance) or giant magnetoresistive arrangements.

These Spin valve magnetoresistance arrangements with tunnel junction or with complete metallic spin valve with current flow perpendicular to the plane of the Allows layers, the spatial Resolution of the Magnetic head along to improve the tracks.

But in terms of sensitivity, these arrangements are not quite satisfactory.

Namely, the tunnel junction devices have too high a resistance level causing a large shot noise. In order to reduce this shot noise to an acceptable level, the product of R.A. must be reduced below 5 Ω.μm ² , this product R.A being the product of the resistance R of the transition times its area A.

The resistance of the spin valve varies exponentially with the thickness of the non-conducting barrier. To obtain a spin valve having a product R.A. which reaches 10 Ω.μm ² , the thickness of its barrier must be only 0.5 nm, which is very little. In addition, the amplitude of the magnetoresistance increases from such assemblies with nonconductive barrier when one attempts to reduce the product R. A below about 10 Ω.μm. ² This is therefore undesirable.

With respect to the fully metallic magnetoresistive spin valve arrangements, they are found to provide a product R.A. which is very small at about 1 to 5 Ω.μm ² . The surfaces of such spin valves amount to a few thousandths to a few hundredths of micron square with information to be read densities of about 150 to 300 Gbit / inch ^2, which leads to resistances of the order of a few tenths of ohms. These values are too low with respect to the resistances of the contacts with the power supply electrodes, to the resistances of the electrodes themselves. The resistances of the electrodes and the contacts are connected in series with that of the spin valve, which has the effect of increasing the signal to be supplied Detect seek to dilute or weaken.

at The working memories have the high resistance of the tunnel junction the disadvantage of leading to a high time constant of the memory, which limits its operating frequency.

The Document EP-A-0 780 912 (best known in the art) an arrangement according to the preamble of claim 1 and a method according to the preamble of claim 17.

PRESENTATION THE INVENTION

The The present invention has the object, a magnetoresistive spin valve arrangement to realize that does not have the disadvantages mentioned above. Yet more specifically, it tries a magnetoresistive spin valve arrangement to realize that has a high magnetoresistance amplitude and thereby with respect to the prior art, a middle product from R.A.

Around To achieve this, the magnetoresistive arrangement comprises a spin valve, which is formed by a stack of layers, among which there are at least two layers in which the relative Orientation of their magnetization directions under the influence of a magnetic field, and facilities that allow in the spin valve to flow a current, transverse or vertical to the level of layers. The spin valve comprises in one of the magnetic Layers at least one non-conductor or semiconductor layer, interrupted of electrically conductive Bridges, which traverse the thickness of the nonconductor or semiconductor layer, these bridges are destined to the current, the transverse or perpendicular through this Stack flows, to concentrate locally.

at this version can the bridges be realized in the magnetic material of the layer, which the non-conductor or semiconductor layer receives.

A continuous non-magnetic, electrically conductive separating layer can be inserted between the non-conductor or semiconductor layer with the electrical conductive bridges and at least one of the magnetic layers. It guarantees a magnetic decoupling of the two magnetic layers.

The electrically conductive bridges can of a non-magnetic material selected between precious metals such as gold, silver, copper or their alloys.

at a variant can They made of a magnetic material such as cobalt, iron, nickel or their alloys.

The non-conductive or semiconducting material of the interrupted layer can be magnetic or non-magnetic. It is not a tunnel crossing. Its role is to redirect the current flowing in the stack, so that he is in the bridges concentrated.

The one of the magnetic layers may have a fixed direction of magnetization have through the connection with an antiferromagnetic layer beyond the magnetic layer with fixed magnetization direction with respect to the non-conductor or semiconductor layer with the electrical conductive Bridges.

At least one of the magnetic layers may pass through a stack of intermediate layers be formed.

Especially the magnetic layer with fixed magnetization direction can formed by a nonmagnetic electrically conductive layer, surrounded by two magnetic intermediate layers.

The discontinuous non-conductor or semiconductor layer may be on oxide, Nitride, semiconductor base can be realized.

The Facilities that enable to flow an electric current to let, can Two electrodes sandwich the spin valve between them contain. Between one of the electrodes and the spin valve can at least one buffer layer are provided.

The Spin valve can be simple or dual.

The The present invention also relates to a reading magnetic head comprising a as defined or as defined above magnetoresistive arrangement includes.

The The present invention also relates to a memory with a memory dot matrix, wherein each of the Memory points as described above or magnetoresistive Arrangement includes.

The present invention also relates to an implementation method of a magnetoresistive spin valve arrangement. It includes the following steps:
Production - for the realization of the spin valve - of a stack of layers with at least two magnetic layers whose orientation concerning their magnetization directions can vary under the influence of a magnetic field,
Manufacturing means for allowing a current to flow in the spin valve transversely to the plane of the layers,
Producing, in one of the magnetic layers, at least one dielectric or semiconductor layer interrupted by electrically conductive bridges passing through the thickness of the layer, these bridges serving to concentrate the current flowing transversely through the stack.

The Non-conductor or semiconductor layer with the electrically conductive bridges can by depositing a thin Layer of non-conductive or semiconductive material with enclosed metallic particles is realized, these being metallic Particles group or regroup to form the bridges.

at a variant, the non-conductor or semiconductor layer with the electrically conductive Realized bridges be by placing an electrically conductive layer, which is located in the Stack is to make it nonconductive or semiconducting, is treated locally. This treatment may be during the deposition or take place at the end of the deposition of the electrically conductive layer.

at In another variant, the non-conductor or semiconductor layer realized with the electrically conductive bridges Be prepared by placing an insulating layer intermittently between two electrically conductive Inserts layers of the stack and by passing the electrically conductive material through the breaks the insulating layer can diffuse. These interruptions in the insulating layer can Naturally be or artificial be generated.

at Another variant realizes the non-conductor or semiconductor layer with the electrically conductive Bridges, by having an electrically conductive Material with low wettability used to drop a variety of drops at its deposition by getting on top of it with the drops not miscible electrically conductive Material separates to the rooms to fill between the drops, and by giving a treatment of the electrically conductive material performs, that is between the drops to make it non-conductive or to make semiconducting.

The Thickness of the electrically conductive Material is slightly smaller than the drops.

The electrically conductive Material that leads to the formation of the non-conductor or semiconductor layer serves, can be deposited on one of the layers of the stack, an electrically conductive Material with low wetting capacity - what to Formation of a variety of drops leads - can be deposited on it and there can be a treatment to do that between the drop of electrically conductive material is non-conductive or semiconducting.

An electrically conductive material with low wettability and an electrically conductive material can be deposited simultaneously on one layer of the stack, these materials being immiscible being treated to make the electrically conductive material non-conductive or semiconducting chen.

One electrically conductive Material with low wettability - what to a variety of drops leads -, can be deposited on one of the layers of the pile, wherein this layer of the stack is treated to render it non-conductive at the surface or semiconducting and their cooperation with the drops to ensure.

The Treatment may be oxidation or nitration.

SUMMARY THE DRAWINGS

The The present invention will be better understood by reading the description from just the explanation serving and by no means limiting embodiments, with reference to the attached drawings.

The 1 (already described) represents a section of a magnetoresistive arrangement according to the prior art;

the 2 FIG. 12 is a sectional view of a magnetoresistive single-pin valve assembly; FIG.

the 3 shows a part of the magnetoresistive arrangement of 2 with regard to the explanation of their operation;

the 4 shows another example of a magnetoresistive arrangement in section;

the 5A . 5B . 5C show further examples of magnetoresistive arrangements, wherein the 5B is according to the invention;

the 6 to 8th show several variants of magnetoresistive dual-spin valve arrangements, wherein the 8th is according to the invention;

the 9 shows a reading magnetic head with a magnetoresistive device;

the 10 shows a memory in which each memory point comprises a magnetoresistive device according to the invention;

the 11A to 11C show realizations of interrupted non-conductive layers with electrically conductive bridges that they traverse.

In In these figures, the various elements are not in for the sake of clarity represented on the same scale.

DETAILED PRESENTATION OF REALIZATION TYPES

The 2 shows an example of a magnetoresistive spin valve assembly according to the invention in a basic configuration. As seen above, a vertical flow spin valve includes at least two magnetic layers in which the relative orientation of their directions of magnetization can vary under the action of a magnetic field. At least one of the magnetic layers has a free direction of magnetization. In order to simplify the diagrams and descriptions, the examples presented below relate to a structure in which one magnetic layer has a free direction of magnetization and the other one fixed. Both magnetic layers could have free magnetization directions and the matching of one of the structures to the other is easily realized.

The arrangement of 2 includes a spin valve 1 , formed by a layer stack with magnetic layers 2 and 3 , of which at least the layer 2 is magnetically fixed and the layer 3 is magnetically free, as well as inside the stack 1 at least one non-conductor or semiconductor layer 4 , crossed by electrically conductive bridges 5 which are transverse to the plane of the magnetic layers 2 . 3 , Preferably, the non-conductor or semiconductor layer is located 4 because of the presence of the bridges 5 is interrupted, between the two magnetic layers 2 and 3 ,

The magnetization of the magnetically fixed layer 2 has only one direction (only one arrow). The magnetization of the magnetically free layer 3 (two opposite arrows) can freely orient in a magnetic field in which it is immersed and which one wishes to detect in the reading magnetic head application. The relative change in the orientation of the magnetization direction of the two layers 2 . 3 leads to a change in the electrical resistance of the spin valve. The magnetoresistive device also includes facilities 6a . 6b that allow current to flow in the spin valve through the stack 1 through which these devices are two electrodes 6a . 6b include the spin valve 1 limit. The electrodes 6a . 6b also serve to pick up the voltage for a measurement of the resistance change. The magnetic layers 2 and 3 are therefore electrically conductive.

The in the non-conductor or semiconductor layer 4 contained conductive bridges 5 have the function in the stack 1 to concentrate on flowing electricity. They connect the two magnetic layers 2 and 3 electric. They maintain the spin of the electrons for the duration of the traversal. The spin diffusion length is greater than the distance separating the two layers.

The non-conductor or semiconductor layer 4 has the main task, one between the two electrically conductive magnetic layers 2 and 3 to ensure localized electrical isolation. It can also do the job of magnetic isolation between the two magnetic layers 2 and 3 have and is non-magnetic in this case. Its thickness can be about 0.2 to 5 nm.

The principle of obtaining a high resistance value while maintaining a large magnetoresistance amplitude with respect to the prior art arrangements is disclosed in U.S.P. 3 shown a particular detail of the spin valve in the 2 shown arrangement shows.

In the 4 , to which reference is also made, an example of an arrangement is shown which can be compared with the 1 serves. Like in the 1 were in the 4 a buffer layer 7 and an antiferromagnetic layer 8th added. The buffer layer 7 For example, it can be tantalum or NiFeCr. The antiferromagnetic layer may be realized on a manganese basis, for example, PtMn, PdPtMn, PtMnCr, IrMn, FeMn, NiMn, RuRhMn.

The perpendicular to the magnetic layers 2 and 3 flowing electricity is concentrated in the electrically conductive bridges 5 , These bridges 5 whose size is small can be compared with potential wells (puits) and point current sources. These bridges 5 can have some nm diameter if the spin valve has an area of a few hundredths of a micron ² . The streamlines that exist in the magnetic layers 2 and 3 arise and the bridges 5 traverse are shown with solid lines.

In the magnetic layers 2 and 3 the isopotential lines, represented by dotted lines, have the shape of concentric half-spheres whose equator is equal to that of the conductive bridges 5 crossed non-conductor or semiconductor layer 4 located.

If in the spin valve the current 1 flows, the current density is at a distance r of a first end 5-1 one of the bridges 5 : J = 1 / 2π · r ²

And the local electric field is: E = ρ _F · J

E = ρ _F · l / 2π · r ² with ρ _F resistivity of the magnetic material of the magnetic layer 2 prevails at the first end 5-1 the bridge 5 ,

The potential difference between the first end 5-1 the bridge 5 and a part very distant from this first end, for example the electrode 6a that is beyond the magnetic layer 2 located at the first end 5-1 the bridge 5 adjacent, is:
V _∞ - V _5-1 = ρ _F · l / 2π · R with R as the radius of the bridge 5 , which in the example is assumed to be cylindrical.

Inside the bridge 5 the streamlines are parallel to the axis of the cylinder. The potential difference between the two ends 5-1 . 5-2 the bridge 5 obtained by:
V _5-1 - V _5-2 = ρ _P · h · l / π · R ² with ρ _P as the bridge's resistivity 5 and h as the height of the bridge 5 ,

In the same way one obtains the potential difference between the second end 5-2 the bridge 5 and a part very distant from this second end, for example the electrode 6b that is beyond the magnetic layer 3 located at the second end 5-2 the bridge 5 adjacent:
V _5-1 - V _-∞ = ρ _{F '} · l / 2π · R with ρ _F' as the resistivity of the magnetic material to the second end 5-2 the bridge 5 adjacent magnetic layer 3 ,

The potential difference between the two electrodes 6a and 6b that limit the spin valve are obtained by: V _∞ - V _-∞ = (ρ _F · L / 2π · R) + (ρ _P · H · l / 2π · R ² ) + (ρ _{F '} · L / 2π · R).

The resistance between the two electrodes 6a and 6b obtained by: R _CPP = (V _∞ - V _-∞ ) / L R _CPP = (ρ _F / 2π · R) + (ρ _P · H / π · R ² ) + (ρ _{F '} / 2π · R).

For comparison, you get in a classic, fully metallic spin valve like the 1 in which the layer separating the magnetic layers is a nonmagnetic electrically conductive layer with the same resistivity (ρ _P ) as that of the material of the bridges, the resistance being given by: R _{CPP '} = (ρ _F · E / π · L ² ) + (ρ _P · L / π · L ² ) + (ρ _{F '} * E '/ 2π · L ² ) where e and e 'thicknesses of the magnetic layers 20 and 30 represent, 1 the thickness of the release layer 10 and L is the radius of the spin valve, which is assumed to be cylindrical.

The increase in resistance associated with the replacement of the interface 10 through the from the electrically conductive bridges 5 interrupted dielectric or semiconductor layer 4 is of the order of magnitude: R _CPP / R _{CPP '} ≈ L ² /he assuming that e, l and e 'are substantially equal.

For a thickness e of the magnetic layer of about 10 nm, a radius L of the spin valve of about 100 nm and a radius R of the bridge 5 of about 2 nm, one can achieve an increase in resistance on the order of 500, which allows the product R. A to reach orders of magnitude of about 1 Ω.μm ² , which is very satisfactory.

The resistance of the bridge 5 is about 30 Ω with the above dimensions.

The amplitude of the magnetoresistance is in the example of 4 higher than in the example of 1 , Namely, the current density, and hence the voltage variation, is maximum in the separation between the magnetic layers at the level of the electrically conductive bridges 5 , but in this environment, the magnetoresistance arises. If the stack also includes further layers, for example the buffer layer 7 and the antiferromagnetic layer 8th , the addition of these layers reduces the effect of the magnetoresistance very little, because the voltage at the terminals of these layers is low: they are crossed by a low current density. On the other hand, in the example of 1 the streamlines always perpendicular to the interfaces and the magnetoresistance is greatly reduced by the addition of these additional layers. The resistance of these layers adds to the resistance of the magnetic layers, which has the effect of diluting the magnetoresistance and reducing its amplitude. An estimate of the effective enhancement of the magnetoresistance resulting from the elimination of this attenuation by the constriction of the current lines due to the presence of the electrically conductive bridges is shown in the following tables:

EXAMPLE OF FIGURE 1

EXAMPLE OF FIGURE 4

This gain can reach at least 10. The resulting resistances may be adjusted by varying the size and density of the electrically conductive bridges, and may, for example, be in the range 0.1 Ω · μm ² to 1 Ω · μm ² .

The bridges 5 can be realized from a non-magnetic material. One can use a noble metal selected from the group comprising gold, silver, copper or their alloys.

In the example of 4 is the discontinuous dielectric or semiconductor layer 4 not magnetic.

The layer 4 may be realized on an oxide basis, for example of alumina, zirconium, strontium titanate (SrTiO ₃ ), magnesia, cobalt oxide, nickel oxide, iron oxide (Fe ₂ O ₃ ) and / or a nitride, such as aluminum nitride, or a semiconductor material, such as germanium, the silicon, the gallium arsenide. The oxides and the nitrides have better magnetoresistive properties than the semiconductor materials.

In one variant, the bridges 5 of an electrically conductive magnetic material, for example of cobalt. This material may be different from that of the magnetic layers 2 and 3 differ. The difference with respect to the previous situation, in which the bridges are of a non-magnetic material, is that between the two magnetic layers 2 and 3 a magnetic coupling exists. When the magnetizations of the two layers 2 and 3 are antiparallel, forming inside the magnetic bridges 5 a wall parallel to the plane of the layers, separating zones of opposite magnetization.

It is possible between the non-conductor or semiconductor layer 4 with their electrically conductive bridges 5 and the magnetic layers 2 and 3 at least one electrically conductive, non-magnetic continuous separating layer 9-1 . 9-2 to introduce magnetic decoupling. In the 5A , which represents this configuration, form the two electrically conductive separation layers 9-1 and 9-2 together with a non-conductor or semiconductor layer 4 with the electrically conductive bridges 5 a sandwich.

The non-conductor or semiconductor material of the discontinuous layer 4 may be magnetic, for example cobalt oxide or iron cobalt oxide or ferromagnetic spinel Fe ₃ O ₄ . It may also be non-magnetic, for example, Al ₂ O ₃ , SiO ₂ , HfO ₂ , Ta ₂ O ₅ .

One would have in the structure of 5A can provide that the electrically conductive material of the bridges 5 that of the separating layers 9-1 . 9-2 is that by interruptions of the layer 4 made of non-conductive material could have diffused.

It is also possible that the material of the bridges 5 that of one of the magnetic layers 2 or 3 is. This special case is in the 5B shown. It is then possible to provide the nonconductor or semiconductor layer with its conductive bridges in the interior of one of the magnetic layers. In the 5B it is only slightly inside the magnetically free layer 3 , on the other magnetic layer 2 facing side.

The two magnetic layers 2 and 3 are separated by at least one continuous separation layer 9 made of non-magnetic, electrically conductive material, for example of copper. This boils down to the fact that the non-conductor or semiconductor layer 4 with their electrically conductive bridges 5 at the interface between the magnetic layer receiving it 3 and the release layer 9 located.

It is possible that one of the magnetic layers is formed by a stack of several intermediate layers. In the 5C it is the magnetically free layer 3 , The intermediate layers 3-1 and 3-2 of different materials enhance the giant magnetoresistive amplitude by enhancing the spin-dependent diffusion contrast and reducing the magnetic fluctuations at the ambient temperature interfaces. The one, called sensitive intermediate layer 3-2 may be, for example, Ni80Fe20 and the other referred to as interfacial magnetic interlayer, for example, Co90Fe10. The broken non-conductive layer 4 with the electrically conductive bridges 5 is located at the interface between the interfacial dopant layer 3-1 and the other magnetic layer 2 but she could also get into the interfacial doping layer 3-1 be included. In general, the non-conductor or semiconductor layer is located 4 advantageously very close to the other magnetic layer 2 ,

In this example is also a magnetic separation layer 9 between the discontinuous dielectric or semiconductor layer 4 with the electrically conductive bridges and the other magnetic layer 2 intended. From the shift 4 It is believed to be magnetic.

In the 8th , Which represents a variant of an inventive arrangement with a pair of spin valves in series, is the non-conductor or semiconductor layer with its electrically conductive bridges in the interior of the magnetically free layer and not on the surface.

However, it is preferable to use the nonconductor or semiconductor layer 4 with their electrically conductive bridges 5 on the side of the magnetically fixed layer 2 so as not to provide the susceptibility to the magnetically free layer 3 must have, affect. This configuration is in the 6 and 7 shown again representing arrangements with a pair of spin valves in series.

One tries the non-conductor or semiconductor layer 4 with their electrically conductive bridges 5 so as close as possible to the interface between the magnetic layers 2 and 3 otherwise, the current concentration effect at the source of magnetoresistance and also the magnetoresistance amplitude would be reduced.

The magnetically fixed layer 2a . 2 B can be formed by a stack of three layers. You can do it like in the 7 with two magnetically defined intermediate layers ( 2a2 . 2a3 ) 2b2 . 2b3 ) realize a non-magnetic electrically conductive intermediate layer 2a1 . 2b1 sandwich. For example, you can use a stack of CoFe / Ru / CoFe. Materials other than ruthenium, for example Ir, Rh, Re, are possible.

The thickness of the middle intermediate layer 2a1 . 2b1 may be between about 0.3 and 1 nm. Thus one realizes a so-called synthetic magnetically fixed layer. Such a synthetic pinned layer makes it possible to increase the pinning field and enhance its magnetic stability.

In the following, several examples of magnetoresistive arrangements with a pair of spin valves in series will be considered, this structure being known as dual-pin valve. Such a dual spin valve may include: a magnetically free middle layer 3 and two magnetically defined layers 2a . 2 B on both sides of the magnetically free layer 3 , The the dual spin valve of the 6 corresponding stack comprises a buffer layer 7 on one of the electrodes, 6b Called base electrode, a first antiferromagnetic attachment layer 8a , a first magnetically fixed layer 2a , a first dielectric or semiconductor layer 4a with electrically conductive bridges 5a , a magnetically free layer 3 , a second dielectric or semiconductor layer 4b , interrupted by electrically conductive bridges 5b , a second magnetically fixed layer 2 B , a second antiferromagnetic attachment layer 8b and finally the second electrode 6a , Called top electrode. In this configuration, the two have non-conductor or semiconductor layers 4a . 4b interrupted by bridges 5a . 5b , the same structure as in the 4 , The magnetically free layer 3 is formed by a stack of intermediate layers, with two magnetic interfacial doping layers 3-2 . 3-3 , which is an actually magnetically free intermediate layer 3-1 sandwich.

As a result, the 7 considered. The the dual spin valve of the 7 corresponding stack comprises a buffer layer 7 on one of the electrodes, 6b Called base electrode, a first antiferromagnetic attachment layer 8a , a first synthetic magnetic layer 2a , a first dielectric or semiconductor layer 4a with electrically conductive bridges 5a , a first nonmagnetic electrically conductive separating layer 9a , a magnetically free layer 3 , a second non-magnetic electrically conductive separation layer 9b , a second dielectric or semiconductor layer 4b , interrupted by electrically conductive bridges 5b , a second synthetic magnetically defined layer 2 B , a second antiferromagnetic attachment layer 8b and finally the second electrode 6a , Called top electrode.

In this configuration, each of the magnetically defined layers 2a . 2 B synthetic and formed by a stack on intermediate layers with a non-magnetic conductive intermediate layer 2a1 . 2b1 sandwiched between two magnetic attachment layers ( 2a2 . 2a3 ) 2b2 . 2b3 ).

The magnetically free layer 3 corresponds to that of 6 with two magnetic interfacial doping interlayers 3-2 . 3-3 , which is an actually magnetically free intermediate layer 3-1 sandwich.

Now let's take a look 8th , The dual-pin valve corresponding stack of 8th includes a buffer layer 7 over one of the electrodes 6b Called base electrode, a first antiferromagnetic attachment layer 8a , a first specified layer 2a , a first separating layer 9a made of non-magnetic electrically conductive material, one by electrically conductive bridges 5 discontinuous non-conductor or semiconductor layer including magnetically free layer 3 , a second non-magnetic electrically conductive separation layer 9b , a second magnetically fixed layer 2 B , a second antiferromagnetic attachment layer 8b and finally the second electrode 6a , Called top electrode.

The 9 Fig. 10 shows an example of a magnetic head including a magnetoresistive device having a spin valve as described above. The electrodes 6a . 6b that enable the electricity 1 supply and tap the voltage V, serve the magnetic shield. The spin valve 1 is in close proximity to a movable magnetic disk 90 on which data is stored in the form of bits. The magnetic shields 6a . 6b allow the spin valve, individual magnetic over between two consecutive bits without interfering with the adjacent transitions. The magnetically fixed layer 2 and the antiferromagnetic attachment layer 8th have an area smaller than that of the non-conductive layer 4 with the electrically conductive bridges 5 who is facing her. They are with non-conductive material 91 wrapped around the broken layer 4 with the electrically conductive bridges 5 extends. This structure allows the zone in which the current flows from the surface of the magnetic data carrier 90 to remove.

In the 10 an inventive memory is shown. It comprises a matrix of memory points, addressable by rows and columns. Each memory point comprises an arrangement or device according to the invention 101 represented by a resistor in series with a switching device 102 in the form of a transistor. Each device according to the invention is provided with an addressing line 103 connected and the switching device with an addressing column 105 , The addressing lines 103 are connected to the outputs of a row addressing circuit 104 connected and the addressing columns 105 with the outputs of a column addressing circuit 106 ,

A process for producing a non-conductor or semiconductor layer having electrically conductive bridges which traverse this layer vertically will now be described. Such a layer is in the 2 to 8th represented by the reference numerals 4 . 5 ,

In the 11A you see a carrier 110 , And for the realization of the bridges, an electrically conductive, only slightly netzungsfähiges material is deposited in a thin layer on this carrier. The carrier may be made of a magnetic material, for example, NiFe, Co, CoFe, particularly in the case where magnetoresistive devices are realized. This material, when deposited on the surface of the carrier, forms a multiplicity of small drops 111 , As is well known, a thin film with high wetting capability spreads on a solid so as to occupy the largest possible area, while a thin film with poor wetting capability occupies only a small area on the solid. This property is used to make the bridges.

It suffices, this material with poor networkability on the non-conductor or semiconductor material 112 Applying the broken layer to apply.

The material used to make the bridges may be silver deposited in the form of thin layers, for example by PVD (for Physical Vapor Deposition). The silver is only slightly wettable at ambient temperature and even less in heat. If you have a very thin layer of silver on the surface of the carrier 110 whose thickness is about 0.2 nm, silver droplets are formed 111 with a diameter of about 1 to 2 nm. Other materials than silver may be used, for example, copper or gold, but where the weak wetting effect is less pronounced.

Around the broken dielectric or semiconductor layer 112 To realize, one can deposit a conductive material that does not interfere with the silver drops 111 which is made non-conductive or semi-conductive by treatment, which treatment has virtually no effect on the silver droplets.

From this point of view or in this case, for example, it is possible to deposit aluminum by means of PVC, with a smaller thickness than the silver pieces 111 , By treatment, such as by oxidation or nitriding, the aluminum becomes dielectric alumina or aluminum nitride. The oxidation may be natural oxidation or oxidation by oxygen plasma or atomic oxygen, etc. Since the aluminum is easily oxidizable and the silver is not at all, because it is a noble metal, only the aluminum oxidizes to form an aluminum oxide layer 112 about 2 nm thick with the bridges 111 receives.

If a magnetoresistive device is to be realized, one can subsequently deposit another magnetic layer on it.

Instead of the silver droplets first 111 on the carrier 110 it is possible to first deposit a thin film of a material that is to the non-conductor or semiconductor layer 112 leads. For example, one can deposit an aluminum layer and then the silver droplets 111 on the thin aluminum layer. By performing a surface oxidation, the aluminum turns around the droplets in alumina 112 and that under the droplets 111 located aluminum 113 oxidizes little or not, as it is protected. A silver drop 111 and the aluminum under this drop 113 thus form a bridge 5 , It could in this latter case the surface of the alumina 112 be rougher than in the previous case.

In another variant, the material forming the non-conductor or semiconductor layer and the electrically conductive material forming the droplets are simultaneously deposited together. This leads to a heterogeneous alloy. When they are on the surface of the carrier 110 is deposited, the droplets form on the surface of the carrier 111 because the two co-deposited materials can not mix. For example, by oxidizing or nitriding the surface, it forms silver droplets, for example 111 around the dielectric or semiconductor layer 112 out.

Another method is to use the electrically conductive material with low wetting capability directly on the support 110 and this carrier is made non-conductive or semiconductive by surface treatment and becomes the discontinuous layer. In the example of the magnetoresistive device according to the invention, the carrier 110 Co- or CoFe-magnetic. It may be related to the 5B described magnetic interfacial doping intermediate layer act. By directly oxidizing this intermediate layer, one obtains a CoO or CoFeO surface, which is in contact with the droplets 111 cooperates. The surface of the carrier 110 it thickens slightly between the silver droplets when it oxidizes.

Around a device according to the invention to produce with at least one spin valve, one separates successively the layers of the stack. For the preparation of the magnetic layers, formed by a stack of intermediate layers or not, the possible separation layers, the buffer layer and one or Several antiferromagnetic layers can all be known to the person skilled in the art Deposition techniques are used. It is in particular to sputtering, ion beam deposition, molecular beam epitaxy or electrolytic metal deposition. The most critical part is the Production of interrupted dielectric or semiconductor layer with the electrically conductive Bridges.

Of course you can apply all methods described above to those interrupted by conductive bridges Non-conductive layer to realize, but there are others possible.

you can be a thin film made of non-conductive or semiconductor material, for example from alumina or silica, the non-magnetic metallic Solid particles (For example, from the group of precious metals such as silver, gold, copper and their alloys) or magnetic of cobalt or its alloys contains. By choosing the carrier, on which the thin film is deposited, warmed, favored the atomic mobility of metal particles regrouping, around the electrically conductive bridges form. You can also perform one or more anneals to said bridges manufacture. In this case, the carrier comprising the thin film is either one of the magnetic layers or a release layer.

The Deposition of the non-conductive thin film can be done by PVD, with a target of dielectric material, or by reactive sputtering a target of a metallic material under oxidizing The atmosphere.

The Solid particles can also applied directly in the form of metallic aggregates or clusters which originate from a precious metal aggregate source.

A Another method can be the local oxidation of a layer of electrical conductive Material to make these locally non-conductive. The non-oxidized Digits form electrically conductive Bridges. This layer may be one of the magnetic layers or a release layer or a layer deposited for this purpose. The oxidation occurs when the layer is deposited or during the Deposition. In this case one interrupts the deposition, leads the Oxidation through and then continues the deposition.

Another method takes advantage of the fact that when an oxide layer is sandwiched intermittently between two metal layers, metal diffuses through the breaks. These breaks are defects or grain boundaries. Of course, the oxide layer may include breaks. In particular, it can be an ultrathin layer of nanocrystalline oxides, for example magnesium oxide (MgO), hafnium oxide (HfO ₂ ), tantalum oxide (Ta ₂ O ₅ ), cobalt oxide (CoO), nickel oxide (NiO), copper oxide (CuO). It is sufficient to deposit the oxide layer on a metallic carrier, for example the magnetic layers 2 . 3 , one of the intermediate magnetic layers 3-1 . 3-2 or a non-magneti rule electrically conductive separation layer 9 , The deposition can be carried out by means of an oxide target or by reactive sputtering, or by deposition of a metal layer and subsequent oxidation. They are covered with a metal layer, no matter if one of the magnetic layers 2 . 3 , one of the intermediate magnetic layers 3-1 . 3-2 or a non-magnetic electrically conductive separation layer 9 , The electrically conductive bridges then form naturally.

The Interruptions in the oxide layer can, instead of becoming natural to be generated by microelectronic techniques, for example Example through a focused ion beam.

Even though several implementations of the present invention in detail can be represented and described, various changes and modifications are made without departing from the scope of the invention to leave.

Claims (29)

Magnetoresistive spin valve device ( 1 ) formed by a stack of layers with at least two magnetic layers ( 2 . 3 ) whose orientation relating to their magnetization directions can vary under the influence of a magnetic field, and at least one dielectric or semiconducting layer ( 4 ) interrupted by electrically conductive bridges ( 5 ), which traverse the thickness of the dielectric or semiconductive layer, these bridges ( 5 ) are intended to locally concentrate the current flowing across the stack, facilities ( 6a . 6b ) to flow in the spin valve a current transverse to the plane of the layers, characterized in that the dielectric or semiconducting layer ( 4 ) with the electrically conductive bridges ( 5 ) in one of the magnetic layers ( 3 ) is arranged. Magnetoresistive device according to claim 1, characterized in that a non-magnetic, electrically conductive continuous separating layer ( 9 ) is inserted between the dielectric or semiconducting layer ( 4 ) with the electrically conductive bridges ( 5 ) and at least one of the magnetic layers ( 2 . 3 ). Magnetoresistive device according to one of claims 1 or 2, characterized in that the bridges ( 5 ) of the material of the magnetic layer ( 3 ), which are the dielectric or semiconductive layer ( 4 ). Magnetoresistive device according to one of claims 1 or 2, characterized in that the electrically conductive bridges ( 5 ) of a non-magnetic material selected from precious metals such as gold, silver, copper or their alloys. Magnetoresistive device according to one of claims 1 to 3, characterized in that the electrically conductive bridges ( 5 ) are made of a magnetic material such as cobalt, iron, nickel or their alloys. Magnetoresistive device according to one of Claims 1 to 5, characterized in that the dielectric or semiconducting material of the discontinuous layer ( 4 ) is magnetic. Magnetoresistive device according to one of Claims 1 to 5, characterized in that the dielectric or semiconducting material of the discontinuous layer ( 4 ) is non-magnetic. Magnetoresistive device according to one of Claims 1 to 7, characterized in that at least one of the magnetic layers ( 2a ) through a stack of intermediate layers ( 2a1 . 2a2 . 2a3 ) is formed. Magnetoresistive device according to one of claims 1 to 8, characterized in that the direction of magnetization of one of the magnetic layers ( 2 ) is determined by the assignment of an antiferromagnetic layer ( 8th ), located beyond the magnetic layer ( 2 ) with fixed magnetization direction with respect to the dielectric or semiconductive layer ( 4 ) with the electrically conductive bridges ( 5 ). Magnetoresistive device according to claim 9, characterized in that the magnetic layer ( 2a ) with a fixed magnetization direction through a non-magnetic, electrically conductive intermediate layer ( 2a1 ) formed between two magnetic intermediate layers ( 2a2 . 2a3 ). Magnetoresistive device according to one of claims 1 to 10, characterized in that the dielectric or semiconductive discontinuous layer ( 4 ) is realized on oxide, nitride, semiconductor basis. Magnetoresistive device according to one of claims 1 to 11, characterized in that the means for flowing an electric current, two electrodes ( 6a . 6b ) comprising the spin valve ( 1 ) sandwiched. Magnetoresistive device according to claim 12, characterized in that between one of the electrodes ( 6b ) and the spin valve ( 1 ) a buffer layer ( 7 ) is provided. Magnetoresistive device according to one of claims 1 to 13, characterized in that the spin valve is dual. Magnetic read head, characterized in that it a magnetoresistive device according to any one of claims 1 to 14 includes. Memory, comprising a matrix of memory points, addressable by lines ( 103 ) and columns ( 105 ), characterized in that each memory point is formed by one having a current switching device ( 102 ) cooperating magnetoresistive device according to one of claims 1 to 14. Method for implementing a magnetoresistive spin valve device ( 1 ), comprising the following steps: realization - for the production of the spin valve - of a stack of layers with at least two magnetic layers ( 2 . 3 ) whose orientation relating to its magnetization directions can vary under the influence of a magnetic field, realization of devices ( 6a . 6b ) for flowing a current transversely to the plane of the layers in the spin valve, characterized in that it comprises a step of realizing at least one dielectric or semiconductive layer ( 4 ) with the layer thickness crossing electrically conductive bridges ( 5 ) in one of the magnetic layers ( 2 . 3 ), these bridges ( 5 ) serve to concentrate the current flowing across the stack. Realization method according to claim 17, characterized in that the dielectric or semiconducting layer ( 4 ) with the electrically conductive bridges ( 5 ) by depositing a thin layer ( 4 ) is made of dielectric or semiconducting material with entrapped metallic particles, these metallic particles grouping or regrouping around the bridges ( 5 ) to build. Realization method according to claim 17, characterized in that the dielectric or semiconducting layer ( 4 ) with the electrically conductive bridges ( 5 ) is realized by an electrically conductive layer ( 9 ) located in the stack to make it dielectric or semiconductive is treated locally. Realization method according to claim 19, characterized in that the treatment during the deposition or at the end of the deposition of the electrically conductive layer ( 9 ) takes place. Realization process according to claim 17, characterized in that the dielectric or semiconducting layer ( 4 ) with the electrically conductive bridges ( 5 ) realized by an insulating layer ( 4 ) with interruptions between two electrically conductive layers ( 9-1 . 9-2 ) of the stack and by passing the electrically conductive maaterial through the breaks in the insulating layer ( 4 ) can diffuse. A method according to claim 21, characterized in that the interruptions in the insulating layer ( 4 ) are natural. A method according to claim 21, characterized in that the interruptions in the insulating layer ( 4 ) are generated artificially. Realization process according to claim 17, characterized in that the dielectric or semiconducting layer ( 4 ) with the electrically conductive bridges ( 5 ) is used by using an electrically conductive material with low wettability in order to deposit a plurality of droplets ( 111 ), by applying to it an electrically non-miscible electrically conductive material ( 112 ) to fill the spaces between the drops, and by having a treatment of the electric conductive material, which is located between the drops to make it dielectric or semiconductive. Realization method according to claim 24, characterized in that the thickness of the electrically conductive material ( 112 ) is slightly smaller than that of the drops ( 111 ). Realization method according to claim 17, characterized in that an electrically conductive material ( 112 ), which serves to form the dielectric or semiconductive layer, on one of the layers ( 110 ) of the stack, an electrically conductive material with low wettability ( 111 ), which results in the formation of a plurality of drops, is deposited thereon, and a treatment is carried out to remove the between the drops ( 111 ) electrically conductive material ( 112 ) to make dielectric or semiconductive. Realization method according to claim 17, characterized in that an electrically conductive material with low wettability ( 111 ) and an electrically conductive material ( 112 ) simultaneously on one layer ( 110 ) of the stack, whereby these materials, which are immiscible, are treated to form the electrically conductive material ( 112 ) to make dielectric or semiconductive. Realization method according to claim 17, characterized in that an electrically conductive material with low wettability ( 111 ) that leads to a multitude of drops ( 111 ), on one of the layers ( 110 ) of the stack, this layer of the stack being treated to make it dielectric or semiconducting at the surface and its cooperation with the drops ( 111 ) to ensure. Implementation method according to one of claims 24 to 28, characterized in that the treatment is an oxidation or a nitration.